Analog Dialogue: A forum for the exchange of circuits, systems, and software for real-world signal processing.
Home     Analog Devices     Feedback     Subscribe     Archives     简体中文     日本語
Advanced Search

Book Revew

The Circuit Designer’s Companion, Third Edition by Peter Wilson, 2012 Elsevier (Newnes), ISBN-13: 978-0080971384

Reviewed by: Jack Memishian, Analog Devices Fellow, October 5, 2012

One aspect of this book deserves mention right up front: it is not written for designers of integrated circuits (the main concern of many Analog Devices engineers). It’s about designing products big enough to see and hold, and about the properties of the many sorts of discrete and (small-scale) integrated components that are used in such designs. Since I spent many decades designing the sort of board-level products that are the book’s main focus (and still do), and since I see some problems with the book—based on that experience—I feel that a review is worthwhile.

The first edition of this book dates from 1990; the author of the first two editions was one Tim Williams. I can’t tell, of course, whether the issues that I see are due to the original author or to his successor.

Companion is well-intentioned. It seeks to fill in a lot of the practical aspects of circuit design that are not taught in the standard EE curriculum, and it partially succeeds—in the sense that it does contain a lot of material in that area which would normally be obtained only by (hard) experience over the course of many years. However, this wisdom is not always presented in a usable form; in some cases it is quite incomplete.

The book’s first chapter demonstrates the presentation issues. This chapter is about grounding, which sounds like a good starting point, but it isn’t. In order to understand the many issues and conflicts of grounding, one needs to have first at least a basic context about systems and their partitioning: that many blocks may have low-level inputs and high-current and/or high-voltage outputs, that ground may serve as a zero-volt reference for signals (often including the input) in single-ended parts of a system and yet also be a return point for high currents—such as filter cap and load currents—that the source and load may have their own ideas about what constitutes “ground,” and much more. The author probably knows all this—some of it is alluded to here and there—but, in the classic manner of “expert syndrome,” most of the background necessary to really understand what is being said about grounding either isn’t stated explicitly at all or is mentioned only in passing. This is unfortunate; while anyone who has been designing systems for a while will have the necessary context, such a person will also already be aware of most of what is being said about grounding, while a beginner with only academic training will have little or no idea why the points made in the chapter are significant—and will have no hooks on which to hang them. The result is likely to be either a dazed skim of the material or an attempt at memorization of rules that should have been internalized by understanding.

There are some worked-out examples of specific cases—which help to give some concreteness; but even here the ball gets dropped: a calculation of voltage drop in a ground path comes out in the hundreds of microvolts, for example, but the rest of the story—showing how big an error this would be in series with a microphone output, for example—or even making the point that it is additive in series with the signal—doesn’t happen. There are diagrams aplenty, but I wonder if it is all that useful to show a number of awful ways to ground an input stage, even given the accompanying text about how you might be forced to use these for (unexplained) cost reasons.

All of that said, there is a lot of good information in the chapter, which will doubtless be valuable in specific situations—ranging all the way from induced currents in a chassis to the proper way to make an electrical connection to aluminum. But for the just-minted graduate, a “Chapter 0” on the architecture of real systems and the practical ranges of currents and voltages for various sorts of signals would have done wonders to set the stage for Chapter 1.

Most of the remaining chapters of the book are much more specific, each devoted to the characteristics of a particular kind of component. But even here the approach taken seems almost academic. For example, obvious catalog properties like tolerances, tempcos and power ratings of carbon composition resistors are presented, along with the useful tidbit that these particular devices are prized for their ability to take large transient power overloads in stride (a thing that can also be discovered in any catalog). However, no mention is made of one of the more well-known drawbacks of carbon comps, their large “excess noise”, which arises in the presence of a current (caused by the granular nature of the resistance material).

In a similar vein, pretty-looking plots of the normal distribution of values are shown for various component types, but there is no mention of the fact that this is frequently not what you get: a population of ±10% resistors may be found to be missing the entire middle, for example ±5%, because those were cherry-picked by the manufacturer and sold for a higher price. Once again, the theoretical aspect is present, but there’s no mention of the practical despite the claims of the book.

A table of chip resistor sizes, translating standard size notation, e.g., 1206, to millimeter dimensions raises a couple of issues. First, the author doesn’t mention (or seem to notice) that the size notation is simply the dimensions of the resistor in mils. Second, the table only appears in the chapter on resistors, so you won’t find that information while researching capacitors (but no problem with inductors, since the existence of chip inductors isn’t even mentioned).

Another example of the “half-a-loaf” level of the book’s component coverage is dielectric absorption (DA). Although there is one paragraph about this effect in the chapter on capacitors, the only mention of specifics is the statement that polystyrene and polypropylene caps have the least DA: there’s nothing at all about which kinds have the worst, or how bad it gets, let alone a table showing the DA of common dielectrics. And in a large chapter on the properties of printed-circuit boards, DA is never mentioned—even though it can be quite severe in glass-epoxy (mysteriously, on thru-hole boards but not on single-sided; I would really have liked to read something about that, having lived through it).

Yet another example of half-information arises in connection with leakage inductance. The author does a good job of describing how a transformer might not have perfect coupling from primary to secondary, stating that this lack of coupling gives rise to “an inductance which is called the leakage inductance” and even venturing that it can cause ringing effects in switching power supplies (would that that were the only effect!). But he never says where this inductance appears in the circuit model of a transformer, an omission that would be even more serious if there were any model given of a transformer (except one for showing how to measure winding capacitances). This is an absurd omission—it reduces the rest of the discussion of the subject of leakage inductance to zero worth.

These examples are all the sort of real-world practical information that one would hope to find in a book with a self-proclaimed “real-world design outlook,” but which is lacking in Companion. Lest anyone assume that these are isolated instances, I submit that any random bungee-jump into the pages (I stop short of saying “void”) of this book will turn up such an oversight; that’s how most of the examples above were found.

Thus Companion does a somewhat academic (and sometimes quite misleading) job of presenting the basic properties of the many discrete and small-scale integrated (SSI) component types covered (I have refrained, here, from examining the chapter on analog ICs – I think I’d be too harsh a judge!). The information given (when correct) would no doubt be of value for the designer fresh from school and (as is so often the case now) without any hobby experience, but it falls short of revealing the “real world.”

Other chapters cover such subjects as electromagnetic interference (EMI), testability, safety and power supplies, subjects which are a lot less hard-edged than the properties of components. Although I am not really qualified to judge these chapters critically, I would venture that the same level of oversights and omissions is here as is seen in the rest of the book, but with much less harmful effect, since each subject is too large for all-inclusive coverage, even in a book-length treatment.

Despite the above, I would recommend this book to relatively inexperienced designers or those who have not previously worked on the design of board-level systems—simply for the sheer volume of information it contains. The caveats are that there are a lot of omissions and half-truths, and to comprehend some of the material you will already need to be well-versed in the area to really understand what is being said. In quite a few areas, updating the book to the present day has apparently not been done at all; much of it is ten to twenty years out of date.

 Open a Dialogue

Question the authors. Share information with your colleagues. Leave feedback for the editors.

What did you think of this article? Was it useful, timely, well written?

Would you like to see more articles on this topic?

Please leave your comments on Facebook

 

Copyright 1995- Analog Devices, Inc. All rights reserved.